3D TEDOR NMR Experiments for the Simultaneous Measurement of

measurement of multiple carbon-nitrogen distances in uniformly 13C,15N-labeled solids. .... plicates accurate measurements of weak 13C-15N dipolar...
0 downloads 0 Views 332KB Size
Published on Web 08/15/2002

3D TEDOR NMR Experiments for the Simultaneous Measurement of Multiple Carbon-Nitrogen Distances in Uniformly 13C,15N-Labeled Solids Christopher P. Jaroniec, Claudiu Filip,† and Robert G. Griffin* Contribution from the Department of Chemistry and Francis Bitter Magnet Laboratory, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139 Received April 1, 2002

Abstract: We describe three-dimensional magic-angle-spinning NMR experiments for the simultaneous measurement of multiple carbon-nitrogen distances in uniformly 13C,15N-labeled solids. The approaches employ transferred echo double resonance (TEDOR) for 13C-15N coherence transfer and 15N and 13C frequency labeling for site-specific resolution, and build on several previous 3D TEDOR techniques. The novel feature of the 3D TEDOR pulse sequences presented here is that they are specifically designed to circumvent the detrimental effects of homonuclear 13C-13C J-couplings on the measurement of weak 13C15 N dipolar couplings. In particular, homonuclear J-couplings lead to two undesirable effects: (i) they generate anti-phase and multiple-quantum (MQ) spin coherences, which lead to spurious cross-peaks and phasetwisted lines in the 2D 15N-13C correlation spectra, and thus degrade the spectral resolution and prohibit the extraction of reliable cross-peak intensities, and (ii) they significantly reduce cross-peak intensities for strongly J-coupled 13C sites (e.g., CO and CR). The first experiment employs z-filter periods to suppress the anti-phase and MQ coherences and generates 2D spectra with purely absorptive peaks for all TEDOR mixing times. The second approach uses band-selective 13C pulses to refocus J-couplings between 13C spins within the selective pulse bandwidth and 13C spins outside the bandwidth. The internuclear distances are extracted by using a simple analytical model, which accounts explicitly for multiple spin-spin couplings contributing to cross-peak buildup. The experiments are demonstrated in two U-13C,15N-labeled peptides, N-acetyl-L-Val-L-Leu (N-ac-VL) and N-formyl-L-Met-L-Leu-L-Phe (N-f-MLF), where 20 and 26 13C-15N distances up to ∼5-6 Å were measured, respectively. Of the measured distances, 10 in N-ac-VL and 13 in N-f-MLF are greater than 3 Å and provide valuable structural constraints.

Introduction

Recent advances in magic-angle-spinning (MAS) solid-state nuclear magnetic resonance (SSNMR) instrumentation and methodology1-3 have greatly extended the applicability of SSNMR from selectively to multiply 13C,15N-labeled peptides and proteins. At moderate magnetic field strengths of 7-14 T (1H Larmor frequencies of 300-600 MHz), complete sequencespecific 13C and 15N resonance assignments have been obtained for several uniformly 13C,15N-labeled microcrystalline peptides4,5 and peptides that self-assemble to form amyloid fibrils,6,7 and partial resonance assignments were determined for the 76* To whom correspondence should be addressed. E-mail: [email protected]. † On leave from National Institute for Research and Development of Isotopic and Molecular Technologies, R-3400 Cluj, Romania. (1) Bennett, A. E.; Griffin, R. G.; Vega, S. In Solid State NMR IV: Methods and Applications of Solid-State NMR; Blumich, B., Ed.; Springer-Verlag: Berlin, 1994; Vol. 33, pp 1-77. (2) Griffin, R. G. Nat. Struct. Biol. 1998, 5, 508-512. (3) Dusold, S.; Sebald, A. Annu. Rep. Nucl. Magn. Reson. Spectrosc. 2000, 41, 185-264. (4) Rienstra, C. M.; Hohwy, M.; Hong, M.; Griffin, R. G. J. Am. Chem. Soc. 2000, 122, 10979-10990. (5) Detken, A.; Hardy, E. H.; Ernst, M.; Kainosho, M.; Kawakami, T.; Aimoto, S.; Meier, B. H. J. Biomol. NMR 2001, 20, 203-221. (6) Balbach, J. J.; Ishii, Y.; Antzutkin, O. N.; Leapman, R. D.; Rizzo, N. W.; Dyda, F.; Reed, J.; Tycko, R. Biochemistry 2000, 39, 13748-13759. 10728

9

J. AM. CHEM. SOC. 2002, 124, 10728-10742

residue human ubiquitin.8,9 At magnetic fields of 17.5-19 T (1H Larmor frequencies of 750-800 MHz), resonance assignments for several uniformly 13C,15N-labeled proteins have been recently presented.10-13 Specifically, partial assignments have been obtained for the 58-residue bovine pancreatic trypsin inhibitor (BPTI)10 and the 150 kDa LH2 light-harvesting membrane-protein complex,11 and nearly complete sequencespecific 13C and 15N12 and nonexchangeable 1H13 assignments have been reported for the 62-residue R-spectrin SH3 domain. Site-specific resonance assignments are a source of valuable structural information in biomolecular solution- and solid-state NMR because the difference between the experimentally (7) Jaroniec, C. P.; MacPhee, C. E.; Astrof, N. S.; Zurdo, J.; Dobson, C. M.; Griffin, R. G. Abstracts of Papers, 223rd National Meeting of the American Chemical Society, Orlando, FL, Spring 2002; American Chemical Society: Washington, DC, 2002. (8) Straus, S. K.; Bremi, T.; Ernst, R. R. J. Biomol. NMR 1998, 12, 39-50. (9) Hong, M. J. Biomol. NMR 1999, 15, 1-14. (10) McDermott, A.; Polenova, T.; Bockmann, A.; Zilm, K. W.; Paulsen, E. K.; Martin, R. W.; Montelione, G. T. J. Biomol. NMR 2000, 16, 209-219. (11) Egorova-Zachernyuk, T. A.; Hollander, J.; Fraser, N.; Gast, P.; Hoff, A. J.; Cogdell, R.; de Groot, H. J. M.; Baldus, M. J. Biomol. NMR 2001, 19, 243-253. (12) Pauli, J.; Baldus, M.; van Rossum, B.; de Groot, H.; Oschkinat, H. ChemBiochem 2001, 2, 272-281. (13) van Rossum, B. J.; Castellani, F.; Rehbein, K.; Pauli, J.; Oschkinat, H. ChemBiochem 2001, 2, 906-914. 10.1021/ja026385y CCC: $22.00 © 2002 American Chemical Society

3D TEDOR NMR

observed isotropic chemical shifts and the corresponding random coil values can be used to predict the conformation of the protein backbone.14-20 Additional information about secondary and tertiary structure in solid-phase peptides and proteins can be obtained from direct measurements of homo- and heteronuclear dipolar couplings, reintroduced during MAS by using dipolar recoupling techniques.1-3 The experiments used for sequential resonance assignments generally exploit the strongest spinspin interactions and are relatively straightforward to implement in uniformly 13C,15N-labeled systems. On the other hand, longrange distance measurements in these systems are more complicated. The nuclei form a tightly coupled network via through-space (dipolar coupling) and through-bond (J-coupling) spin-spin interactions; consequently, the strongest couplings can interfere with the accurate determination of the weaker ones.21-25 In the following, we will focus our attention on the problem of heteronuclear 13C-15N recoupling in multiple-spin systems, with the ultimate goal of measuring multiple carbonnitrogen distances simultaneously in uniformly 13C,15N-labeled peptides and proteins. A number of experiments have been developed to reintroduce the dipolar coupling between a pair of low-γ spin I ) 1/2 nuclei (e.g., 13C, 15N, 31P) during MAS.1-3 Some of the most robust and versatile heteronuclear recoupling experiments are derived from rotational echo double resonance (REDOR)26,27 and the closely related transferred echo double resonance (TEDOR).28,29 (Applications of these methods to systems of importance in biology and materials science have been summarized in several recent reviews.1-3,30) During REDOR,26,27 the I-S dipolar coupling is reintroduced by a train of rotor-synchronized 180° pulses applied to the S-spins. This results in the dephasing of the I-spin-echo with a modulation frequency proportional to the magnitude of the dipolar coupling. TEDOR28,29 (the basic pulse sequence consists of two REDOR periods bracketing a pair of 90° pulses on I- and S-spins) acts to transfer magnetization between dipolar coupled I and S nuclei, where the buildup of the S-spin magnetization is proportional to the magnitude of the I-S dipolar coupling. TEDOR coherence transfer dynamics are analogous to the well-known solution-state INEPT experiment,31 which uses the heteronuclear J-coupling to achieve coherence transfer, and solid-state INEPT experiments,32 for which coherence transfer is accomplished via heteronuclear dipolar couplings. While REDOR and TEDOR are both well(14) Spera, S.; Bax, A. J. Am. Chem. Soc. 1991, 113, 5490-5492. (15) Wishart, D. S.; Sykes, B. D. Methods Enzymol. 1994, 239, 363-392. (16) Wishart, D. S.; Bigam, C. G.; Holm, A.; Hodges, R. S.; Sykes, B. D. J. Biomol. NMR 1995, 5, 67-81. (17) Cornilescu, G.; Delaglio, F.; Bax, A. J. Biomol. NMR 1999, 13, 289-302. (18) Saito, H. Magn. Reson. Chem. 1986, 24, 835-852. (19) Saito, H.; Tuzi, S.; Naito, A. Annu. Rep. Nucl. Magn. Reson. Spectrosc. 1998, 36, 79-121. (20) Luca, S.; Filippov, D. V.; van Boom, J. H.; Oschkinat, H.; de Groot, H. J. M.; Baldus, M. J. Biomol. NMR 2001, 20, 325-331. (21) Costa, P. R. Ph.D. Thesis, Massachusetts Institute of Technology, 1996. (22) Hohwy, M.; Rienstra, C. M.; Jaroniec, C. P.; Griffin, R. G. J. Chem. Phys. 1999, 110, 7983-7992. (23) Ladizhansky, V.; Vega, S. J. Chem. Phys. 2000, 112, 7158-7168. (24) Fyfe, C. A.; Lewis, A. R. J. Phys. Chem. B 2000, 104, 48-55. (25) Hodgkinson, P.; Emsley, L. J. Magn. Reson. 1999, 139, 46-59. (26) Gullion, T.; Schaefer, J. J. Magn. Reson. 1989, 81, 196-200. (27) Gullion, T.; Schaefer, J. AdV. Magn. Reson. 1989, 13, 57-83. (28) Hing, A. W.; Vega, S.; Schaefer, J. J. Magn. Reson. 1992, 96, 205-209. (29) Hing, A. W.; Vega, S.; Schaefer, J. J. Magn. Reson. A 1993, 103, 151162. (30) McDowell, L. M.; Schaefer, J. Curr. Opin. Struct. Biol. 1996, 6, 624629. (31) Morris, G. A.; Freeman, R. J. Am. Chem. Soc. 1979, 101, 760-762. (32) Roberts, J. E.; Vega, S.; Griffin, R. G. J. Am. Chem. Soc. 1984, 106, 25062512.

ARTICLES

established for selectively 13C,15N-labeled samples, their application to multiply labeled systems is less straightforward. Depending on the details of the particular experiment, the recoupling dynamics can depend on multiple spin-spin couplings and their relative orientations, which potentially complicates accurate measurements of weak 13C-15N dipolar couplings.24 Various experiments based on REDOR and TEDOR have been proposed to circumvent the problems associated with multiple-spin systems.33-46 Most recently, we described a frequency-selective REDOR (FSR)43 experiment and demonstrated its applications to uniformly 13C,15N-labeled peptides and proteins.43,47 The experiment was specifically developed to perform accurate site-specific measurements of weak 13C15N dipolar couplings in U-13C,15N-labeled samples. Accordingly, for a tripeptide N-formyl-L-Met-L-Leu-L-Phe (N-f-MLF), 16 13C-15N distances up to ∼6 Å were measured by using FSR with ∼0.1-0.3 Å precision43 and used in combination with multiple dihedral angle constraints48 to calculate the threedimensional high-resolution structure of N-f-MLF.49 In the special case of a biological macromolecule exhibiting sufficient 13C and 15N chemical shift resolution, frequency-selective REDOR was applied to a 248-residue membrane protein, bacteriorhodopsin, to determine two distances between the retinal Schiff base nitrogen and carboxyl groups of aspartate residues, D85 and D212, across the active site.47 In the following we discuss three-dimensional heteronuclear correlation (HETCOR) experiments, which are designed to simultaneously measure multiple long-range carbon-nitrogen distances in uniformly 13C,15N-labeled samples. During a 3D HETCOR experiment, a series of 2D I-S chemical shift correlation spectra are recorded as a function of the duration of the I-S coherence transfer period (also referred to as the mixing time); for distance measurements the coherence transfer proceeds via I-S dipolar couplings. Cross-peaks located at (ΩSj,ΩIi) in the 2D spectra, where ΩIi and ΩSj are the isotropic chemical shifts of the ith I-spin and the jth S-spin, respectively, are observed only for I-S spin pairs having a sufficiently strong dipolar coupling, and the buildup of cross-peak intensities as a function of the mixing time provides information about I-S distances in a site-specific fashion. The experiments discussed here employ TEDOR for 13C-15N coherence transfer and build (33) Fyfe, C. A.; Mueller, K. T.; Grondey, H.; Wong-Moon, K. C. Chem. Phys. Lett. 1992, 199, 198-204. (34) van Eck, E. R. H.; Veeman, W. S. J. Magn. Reson. A 1994, 109, 250252. (35) Michal, C. A.; Jelinski, L. W. J. Am. Chem. Soc. 1997, 119, 9059-9060. (36) Bennett, A. E.; Becerra, L. R.; Griffin, R. G. J. Chem. Phys. 1994, 100, 812-814. (37) Bennett, A. E.; Rienstra, C. M.; Lansbury, P. T.; Griffin, R. G. J. Chem. Phys. 1996, 105, 10289-10299. (38) Schaefer, J. J. Magn. Reson. 1999, 137, 272-275. (39) Gullion, T.; Pennington, C. H. Chem. Phys. Lett. 1998, 290, 88-93. (40) Liivak, O.; Zax, D. B. J. Chem. Phys. 2000, 113, 1088-1096. (41) Liivak, O.; Zax, D. B. J. Chem. Phys. 2001, 115, 402-409. (42) Jaroniec, C. P.; Tounge, B. A.; Rienstra, C. M.; Herzfeld, J.; Griffin, R. G. J. Am. Chem. Soc. 1999, 121, 10237-10238. (43) Jaroniec, C. P.; Tounge, B. A.; Herzfeld, J.; Griffin, R. G. J. Am. Chem. Soc. 2001, 123, 3507-3519. (44) Chan, J. C. C.; Eckert, H. J. Magn. Reson. 2000, 147, 170-178. (45) Chan, J. C. C. Chem. Phys. Lett. 2001, 335, 289-297. (46) Chan, J. C. C.; Eckert, H. J. Chem. Phys. 2001, 115, 6095-6105. (47) Jaroniec, C. P.; Lansing, J. C.; Tounge, B. A.; Belenky, M.; Herzfeld, J.; Griffin, R. G. J. Am. Chem. Soc. 2001, 123, 12929-12930. (48) Rienstra, C. M.; Hohwy, M.; Mueller, L. J.; Jaroniec, C. P.; Reif, B.; Griffin, R. G. J. Am. Chem. Soc. 2002, in press. (49) Rienstra, C. M.; Tucker-Kellogg, L.; Jaroniec, C. P.; Hohwy, M.; Reif, B.; McMahon, M. T.; Tidor, B.; Lozano-Perez, T.; Griffin, R. G. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 10260-10265. J. AM. CHEM. SOC.

9

VOL. 124, NO. 36, 2002 10729

ARTICLES

Jaroniec et al.

on 3D TEDOR techniques described previously by Fyfe and co-workers,33 van Eck and Veeman,34 and Michal and Jelinski.35 We note that a 3D HETCOR experiment of this type based on band-selective 13C-15N cross-polarization4,50 has also been discussed.51 The previous 3D TEDOR pulse sequences,33,34 where the initial S-spin magnetization is frequency labeled and transferred by using TEDOR to the I-spins for detection, are not directly applicable to uniformly 13C,15N-labeled systems, because 13C13C dipolar and/or 13C and 15N chemical shift anisotropy (CSA) interactions are active during the mixing period and interfere with the 13C-15N coherence transfer process. These 3D TEDOR experiments were recently extended to multiple 13C,15N spin systems by Michal and Jelinski,35 who described an approach where the initial 13C magnetization is transferred to 15N for frequency labeling and subsequently transferred back to 13C for detection; the resulting experiment is relatively insensitive to strong 13C-13C dipolar couplings and significant 13C and 15N CSA interactions. While this modified 3D TEDOR experiment35 addresses many of the problems associated with uniformly 13C,15N-labeled samples, it is not explicitly compensated for homonuclear 13C-13C J-couplings (J ) 30-60 Hz in peptides52), which interfere with the measurement of weak 13C15N dipolar couplings. These homonuclear J-couplings generate anti-phase and multiple-quantum (MQ) spin coherences during the mixing period, which lead to spurious cross-peaks and phasetwisted lines in the 2D spectra, and prohibit the extraction of reliable cross-peak intensities. These problems become particularly severe for mixing times on the order of tmix ≈ 1/(2J). Furthermore, the cross-peaks are modulated by the J-evolution, which leads to significantly reduced intensities for strongly J-coupled 13C sites (e.g., CO and CR). In this paper, we describe two modified 3D TEDOR experiments specifically designed to address the problems associated with the homonuclear J-couplings. The first approach employs z-filter periods to suppress the anti-phase and MQ coherences, generates 2D 15N-13C correlation spectra with pure absorption mode peaks for all TEDOR mixing times, and provides information about all carbon-nitrogen distances simultaneously. The second approach uses band-selective 13C pulses to refocus J-couplings between 13C spins within the selective pulse bandwidth and 13C spins outside the bandwidth and is particularly applicable to distance measurements involving strongly J-coupled 13C sites (homonuclear J-decoupling of this type has been used previously in solution- and solid-state NMR experiments42,43,53,54). The 13C-15N distances are extracted by using a simple analytical model, which accounts explicitly for multiple spin-spin interactions contributing to cross-peak buildup. The methods are demonstrated in model U-13C,15N-labeled peptides, N-acetyl-L-Val-L-Leu and N-formyl-L-Met-L-Leu-L-Phe, where 20 and 26 13C-15N distances up to ∼5-6 Å were measured, respectively. (50) Baldus, M.; Petkova, A. T.; Herzfeld, J.; Griffin, R. G. Mol. Phys. 1998, 95, 1197-1207. (51) Rienstra, C. M. Presented at the 2nd Alpine Conference on Solid-State NMR, Chamonix, France, 2001. (52) Cavanagh, J.; Fairbrother, W. J.; Palmer, A. G.; Skelton, N. J. Protein NMR Spectroscopy: Principles and Practice; Academic Press: San Diego, 1996. (53) Bru¨schweiler, R.; Griesinger, C.; Sørensen, O. W.; Ernst, R. R. J. Magn. Reson. 1988, 78, 178-185. (54) Straus, S. K.; Bremi, T.; Ernst, R. R. Chem. Phys. Lett. 1996, 262, 709715. 10730 J. AM. CHEM. SOC.

9

VOL. 124, NO. 36, 2002

Figure 1. 3D z-filtered TEDOR (a) and 3D band-selective TEDOR (b) pulse sequences. Narrow and wide solid rectangles represent π/2 and π pulses, respectively. During REDOR, two π pulses per rotor period are applied on the 15N channel and phase cycled according to the xy-4 scheme.56 The short delay, τ, ensures that the total delay between the REDOR mixing periods is equal to an integer number of rotor cycles and is required for efficient reconversion of anti-phase coherences into observable 13C magnetization. In (a), the modifications to the recently described 3D TEDOR experiment35 are highlighted by gray rectangles. The modifications consist of two z-filter periods, ∆, which eliminate multiple-quantum and anti-phase spin coherences generated by 13C-13C J-evolution (see text for details). During the z-filters, a weak proton rf field, ωrf ≈ ωr, was applied57 to facilitate the rapid dephasing of transverse 13C spin coherences via the contact with the abundant proton bath. In (b), band-selective 13C Gaussian π pulses refocus 13C-13C J-couplings to nuclei outside the π pulse bandwidth. The phase cycles were as follow: (a) φ1 ) 16×(1) 16×(3), φ2 ) 1, φ3 ) 13, φ4 ) 2244, φ5 ) 11112222 33334444, φrec’r ) 42241331 24423113 24423113 42241331; (b) φ1 ) 8×(1) 8×(3), φ2 ) 1, φ3 ) 13, φ4 ) 11223344, φrec’r ) 31241342 13423124, where 1 ) x, 2 ) y, 3 ) -x, 4 ) -y. All remaining pulses (except 15N π pulses during REDOR which were phase cycled xy - 4 and 1H spin-lock which had phase y) have phase x. Hypercomplex data were acquired by shifting φ3 according to Ruben and co-workers.58

Experimental Section NMR Experiments. NMR spectra were recorded at 11.7 T (500.1 MHz 1H, 125.8 MHz 13C, 50.7 MHz 15N) by using a custom-designed spectrometer (courtesy of Dr. David J. Ruben) with a Chemagnetics (Fort Collins, CO) triple-resonance MAS probe. The probe was equipped with a 4.0 mm Chemagnetics spinning module. The spinning frequency of 10.0 kHz was used in the experiments and regulated to (5 Hz with a Doty Scientific (Columbia, SC) spinning frequency controller. Samples were centerpacked in the rotors to minimize the effects of rf inhomogeneity. The 3D TEDOR pulse sequences are shown in Figure 1 and described in detail in the Theoretical Background section below. Unless otherwise indicated, the 15N REDOR π pulse length was 10 µs, the 13C and 15N π/2 pulses were 5 µs, the TPPM decoupling was ∼100 kHz (total phase difference, 12°; TPPM pulse length, 5.0 µs) during mixing and ∼83 kHz (total phase difference, 12°; TPPM pulse length, 6.0 µs) during t1 and t2, the CW decoupling was 100 kHz, and the 13C band-selective Gaussian pulse (Figure 1b) was 0.4 ms, divided into 64 increments, and truncated at 1% of the maximum amplitude. U-13C,15N-Labeled Peptides. The peptides used in the experiments were [U-13C,15N]N-acetyl-L-Val-L-Leu (N-ac-VL) and N-formyl-[U13 15 C, N]L-Met-L-Leu-L-Phe (N-f-MLF). The U-13C,15N-labeled amino acids and [1,2-13C]acetic anhydride used in the peptide synthesis were purchased from Cambridge Isotope Laboratories (Andover, MA). N-acVL and N-f-MLF were synthesized by Synpep (Dublin, CA) and American Peptide (Sunnyvale, CA), respectively, using standard solidphase methods and purified by HPLC. To minimize the effects of intermolecular couplings on the measured distances, the U-13C,15Nlabeled peptides were diluted in the respective natural abundance peptides followed by recrystallization from appropriate solvents: N-ac-

3D TEDOR NMR

ARTICLES

VL was diluted to ∼20% and crystallized from a 1:1 (v/v) H2O:acetone solution,55 and N-f-MLF was diluted to ∼10% and crystallized from 2-propanol.4,43

Theoretical Background

3D TEDOR Pulse Sequences. The 3D TEDOR experiments are presented in Figure 1. The 3D z-filtered TEDOR sequence (Figure 1a) (also denoted in the following as 3D ZF TEDOR) provides information about all 13C-15N dipolar couplings simultaneously. The 3D band-selective TEDOR sequence (3D BASE TEDOR) shown in Figure 1b probes dipolar couplings for 13C spins within the selective pulse bandwidth and is particularly applicable to distance measurements involving sites with strong 13C-13C J-couplings (e.g., CO and CR). The desired TEDOR coherence transfer pathway for a heteronuclear I-S spin pair (I and S represent 13C and 15N, respectively) with the effective dipolar coupling, ω, is 90x(I),90x(S)

REDOR

Ix 98 2IySz sin(ωtmix/2) 98 t1

-2IzSy sin(ωtmix/2) 98 90-x(I),90x/y(S)

- 2IzSy sin(ωtmix/2) eiΩSt1 98 REDOR

-2IySz sin(ωtmix/2) eiΩSt1 98 t2

Ix sin2(ωtmix/2) eiΩSt1 98 Ix sin2(ωtmix/2) eiΩSt1 eiΩIt2 The transverse I-spin (13C) magnetization is created by using ramped cross-polarization from protons.59 Subsequently, a REDOR sequence,26,27 initiated for a time tmix/2, reintroduces the I-S dipolar coupling and generates I-spin coherence in antiphase with respect to the S-spin (∼IySz). The first pair of 90° pulses on the I- and S-spins results in coherence transfer to the S-spin (∼IzSy). This coherence is frequency labeled in t1 with the S-spin chemical shift, ΩS. The second pair of 90° pulses on the I- and S-spins transfer the coherence back to the I-spin (∼IySz). The resulting anti-phase coherence is reconverted into observable I-spin magnetization (∼Ix) during the second REDOR period and frequency labeled with the I-spin chemical shift, ΩI, during the acquisition of the FID in t2. Assuming quadrature detection in t1 and t2,52,60 a Fourier transformation of the time-domain data results in a 2D spectrum with a crosspeak at the frequency (ΩS,ΩI). For each crystallite in the powder sample, the cross-peak intensity is modulated as a function of the mixing time, tmix, according to sin2(ωtmix/2). In our implementation of 3D TEDOR experiments for uniformly 13C,15N-labeled samples, all REDOR pulses are applied on the 15N channel to avoid the recoupling of 13C spins61,62 and phased according to the xy-4 scheme56 to compensate for pulse imperfections and resonance offset effects (55) Stewart, P. L.; Tycko, R.; Opella, S. J. J. Chem. Soc., Faraday Trans. 1 1988, 84, 3803-3819. (56) Gullion, T.; Baker, D. B.; Conradi, M. S. J. Magn. Reson. 1990, 89, 479484. (57) Oas, T. G.; Griffin, R. G.; Levitt, M. H. J. Chem. Phys. 1988, 89, 692695. (58) States, D. J.; Haberkorn, R. A.; Ruben, D. J. J. Magn. Reson. 1982, 48, 286-292. (59) Metz, G.; Wu, X.; Smith, S. O. J. Magn. Reson. A 1994, 110, 219-227. (60) Ernst, R. R.; Bodenhausen, G.; Wokaun, A. Principles of Nuclear Magnetic Resonance in One and Two Dimensions; Clarendon Press: Oxford, 1991. (61) Gullion, T.; Vega, S. Chem. Phys. Lett. 1992, 194, 423-428. (62) Bennett, A. E.; Ok, J. H.; Griffin, R. G.; Vega, S. J. Chem. Phys. 1992, 96, 8624-8627.

(pulse timings have been described in detail elsewhere43). The MQ and anti-phase terms in the density matrix created by homonuclear J-evolution are suppressed by using two z-filter periods, ∆, highlighted by gray rectangles. For 3D BASE TEDOR, the band-selective 13C pulse, centered in the period occupying an even number of rotor cycles (2mτr), refocuses homonuclear J-couplings. The influence of proton couplings during the 13C-15N recoupling periods is attenuated by a combination of continuous-wave (CW) and two-pulse phase modulation (TPPM)63 decoupling, as described in detail previously.43 TPPM decoupling was used during t1 and t2. The delay, τ, following t1 evolution ensures that the total delay between the REDOR mixing periods is equal to an integral number of rotor cycles and is crucial to the efficient reconversion of antiphase coherences to observable 13C magnetization. Spin Dynamics in U-13C,15N-Labeled Systems. We consider a uniformly 13C,15N-labeled sample spinning rapidly about the magic angle (θm ) tan-1 x2) at a frequency far from rotational resonance conditions.64 For the 3D TEDOR pulse sequence shown in Figure 1a, the spin system can be considered to evolve under isotropic chemical shifts and homonuclear J-couplings during the free evolution periods, t1 and t2,

H)

πJjk2IjzIkz ∑k ΩIkIkz + ∑k ΩSkSkz + ∑ j ∼3 Å) were measured, and in N-f-MLF, 13 such distances were determined. A number of distances in the 3-4 Å regime (e.g., Val(Cβ)-Leu(N), Val(Cγ1)-Val(N), and Leu(Cγ)Leu(N) in N-ac-VL; Met(Cβ)-Leu(N), Met(Cγ)-Met(N), and Leu(Cβ)-Phe(N) in N-f-MLF) depend on a single ψ or χ torsion J. AM. CHEM. SOC.

9

VOL. 124, NO. 36, 2002 10739

ARTICLES

Jaroniec et al.

Figure 12. Representative two-dimensional slices from the 3D BASE TEDOR experiment for 13CO nuclei and cross-peak buildup curves in N-formyl[U-13C,15N]L-Met-L-Leu-L-Phe. The pulse sequence in Figure 1b was used with the carrier frequency on resonance with the carbonyl region of the spectrum and a Gausian band-selective 13C refocusing pulse of 0.4 ms. Experimental parameters are described in the text. The 3D data set was recorded as (12, 1024, 8) points in (t1, t2, tmix), with time increments of (0.8, 0.02, 2.0) ms, resulting in acquisition times of (8.8, 20.5, 16.0) ms. Each FID was 256 scans, the recycle delay was 3 s, and (t1, t2) points were acquired as complex pairs, yielding a total measurement time of ∼42 h. The buildup curves for Met C′ (c) and Leu C′ (d) corresponding to the Met (solid circles), Leu (gray circles), and Phe (open circles) 15N sites were fit (s) by using analytical expressions in eq 10, and the distances determined are indicated in the insets. All 13C-15N internuclear distances measured in N-f-MLF are summarized in Table 3.

calculation of a three-dimensional SSNMR structure for the tripeptide.49 Conclusions

Figure 13. Comparison of carbon-nitrogen distances in N-formyl-L-MetL-Leu-L-Phe measured by using 3D ZF TEDOR (b) and 3D BASE TEDOR (O) with the corresponding X-ray diffraction distances determined for N-f-MLF-OMe70 (see text for discussion of uncertainties in measured distances).

angle, and the longer distances in the 4-6 Å range contain information about multiple φ, ψ, and χ torsion angles. These 13C-15N distance measurements place very strong constraints on the overall peptide structure (particularly on the side chain conformations), but in general they should be used in combination with other structure determination methods, such as chemical shift assignments, and direct measurements of torsion angles, to uniquely define a three-dimensional peptide structure. This approach was used for N-f-MLF, where long-range 13C-15N distance measurements43 were combined with measurements of 10 φ, ψ, and χ torsion angles from dipolar tensor correlation experiments48 and used as constraints in the 10740 J. AM. CHEM. SOC.

9

VOL. 124, NO. 36, 2002

In conclusion, we have presented two 3D TEDOR NMR experiments for the simultaneous measurement of multiple carbon-nitrogen distances in uniformly 13C,15N-labeled solids in site-specific fashion. The novel feature of the experiments is that they are specifically compensated for the detrimental effects of 13C-13C J-couplings on the measurements of long-range carbon-nitrogen distances. The experiments are robust and straightforward to implement. Qualitatively, the 13C-15N correlations in themselves provide direct information about threedimensional structure, and quantitative distance measurements were performed by using a simple analytical simulation model, which accounts explicitly for multiple spin-spin couplings contributing to cross-peak buildup. The utility of the experiments for structural studies was demonstrated by measurements of 20 and 26 carbon-nitrogen distances up to ∼5-6 Å in two uniformly 13C,15N-labeled peptides. Specifically, 10 13C-15N distances measured in N-ac-VL and 13 distances in N-f-MLF are greater than 3 Å and provide valuable structural constraints. We expect the 3D TEDOR experiments demonstrated here on microcrystalline peptides to be applicable to structural measurements in noncrystalline biological systems, such as amyloid fibrils and membrane proteins. We have recently used 3D ZF TEDOR to simultaneously measure multiple 13C-15N distances in the 3-5 Å range in a U-13C,15N-labeled amyloid fibrilforming peptide, and applications to distance measurements in the bacteriorhodopsin active site47 can be envisaged.

3D TEDOR NMR

ARTICLES

Acknowledgment. The authors thank Dr. Vladimir Ladizhansky, Dr. Chad Rienstra, Dr. David Ruben, and Dr. Jonathan Lansing for stimulating discussions during the course of this work, and Dr. Bernd Reif for a careful reading of the manuscript. C.P.J. thanks the NSF for a Predoctoral Fellowship. This research was supported by NIH Grants GM-23403, GM-23289, and RR-00995.

I1x f -2I1zSysω1cJ cos(ΩSt1) - ZQxSxsω1sJ{cos[(ΩI1 - ΩI2 + ΩS)t1] cos[(ΩI1 - ΩI2 - ΩS)t1]} + ZQySysω1sJ{cos[(ΩI1 - ΩI2 + ΩS)t1] + cos[(ΩI1 - ΩI2 - ΩS)t1]} + DQxSxsω1sJ{cos[(ΩI1 + ΩI2 + ΩS)t1] cos[(ΩI1 + ΩI2 - ΩS)t1]}

Appendix

In this section we calculate the 3D TEDOR spin dynamics in the simplest multiple-spin system (I2-S) having heteronuclear dipolar and homonuclear J-couplings. Although this is much simpler than the spin systems represented by the uniformly 13C,15N-labeled peptides, it is sufficient to demonstrate the detrimental effects due to homonuclear J-couplings and explain the main features observed in the experimental 2D correlation spectra shown in Figure 5a-c. We consider the 3D TEDOR pulse sequence shown in the Figure 1a, where the z-filter periods, ∆, are set to zero (i.e., the homonuclear J-coupling is active throughout the entire experiment). In the analysis, it is sufficient to follow the evolution of the transverse magnetization corresponding to one of the spins, for instance I1x. The results can be extended to the second spin by simply interchanging the labels 1 and 2. For consistency, in the following derivations we use the notation introduced in eqs 1-5 of the Theoretical Background section. For the model I2-S spin system, the effective Hamiltonian during the REDOR mixing periods is given by

H)



ωkIkzSz + πJ2I1zI2z

(A1)

k)1,2

Therefore, after the first REDOR mixing period followed by the 90° pulses applied on the I and S channels, I1x transforms to

I1x f (I1xcJ - 2I1zI2ysJ)cω1 - 2(I1zcJ + 2I1xI2ysJ)Sysω1 (A2) The applied phase cycle eliminates all coherences which contain only I-spin operators after the first REDOR mixing period, and retains terms which depend on transverse S-spin operators:

I1x f -2(I1zcJ + 2I1xI2ysJ)Sysω1

(A3)

Thus, in addition to the desired anti-phase coherence (∼I1zSy), which reflects I-S dipolar correlations and carries I-S distance information, an additional term (∼I1xI2ySy) is also generated, which depends on I-I correlations and is represented by a combination of zero-quantum (ZQ) and double-quantum (DQ) spin coherences with respect to I-spin operators. The REDOR excitation is followed by a free evolution period (t1) under the Hamiltonian

H)



k)1,2

This leads to

ΩIkIkz + ΩSSz + πJ2I1zI2z

(A4)

+ DQySysω1sJ{cos[(ΩI1 + ΩI2 + ΩS)t1] + cos[(ΩI1 + ΩI2 - ΩS)t1]} (A5) where we have retained only the cosine-modulated chemical shift terms, and the following notation for the ZQ and DQ coherences was used:52

ZQx ) I1xI2x + I1yI2y,

ZQy ) I1yI2x - I1xI2y

DQx ) I1xI2x - I1yI2y,

DQy ) I1xI2y + I1yI2x

(A6)

The I1zSy coherence in eq A5 is modulated by the S-spin chemical shift, ΩS, and gives rise to the desired cross-peaks located at (ΩS,ΩI) in the 2D spectra. The zero- and doublequantum coherences evolve in the indirect dimension with frequencies Ω1I(Ω2I(ΩS, giving rise to spurious cross-peaks. For the simple I2-S spin system, only the isotropic chemical shift terms in the Hamiltonian in eq A4 contribute to the spin dynamics during t1. For larger spin systems having multiple I-I J-couplings, the scalar interactions would also contribute to the evolution, resulting in more complicated dynamics. The second pair of 90° pulses followed by the REDOR mixing period generates I-spin coherences, which are subsequently detected during the acquisition of the FID in t2. The coherences are represented by both in-phase (∼Ix) and antiphase (∼IyIz) terms: 2 cos(ΩSt1) I1x f - (I1xc2J - 2I1yI2zsJcJ)sω1 2 {cos[(ΩI1 - ΩI2 + ΩS)t1] + - (I1xc2J + 2I1yI2zsJcJ)sω1 cos[(ΩI1 - ΩI2 - ΩS)t1]}

- (I2xc2J + 2I2yI1zsJcJ)sω1sω2{cos[(ΩI1 - ΩI2 + ΩS)t1] - cos[(ΩI1 - ΩI2 - ΩS)t1]} (A7) Under the chemical shift and homonuclear J-coupling interactions, the in-phase terms (∼Ix) in eq A7 give rise to purely absorptive doublets split by the J-coupling,

I1x f I1+(ei(ΩI1-J/2)t2 + ei(ΩI1+J/2)t2)

(A8)

while the anti-phase terms (∼IyIz) generate purely dispersive doublets with anti-phase components:

2I1yI2z f iI1+(ei(ΩI1-J/2)t2 - ei(ΩI1+J/2)t2)

(A9)

The superposition of these signals results in phase-twisted lines in the direct dimension and prohibits the extraction of reliable cross-peak intensities. In summary, even for relatively simple multiple-spin systems, the presence of homonuclear J-couplings leads to quite comJ. AM. CHEM. SOC.

9

VOL. 124, NO. 36, 2002 10741

ARTICLES

Jaroniec et al.

plicated spin dynamics. The J-couplings generate artifacts in the 2D correlation spectra in the form of spurious cross-peaks and phase-twisted lines, which degrade the resolution and prohibit the extraction of reliable cross-peak intensities. The J-coupling effects become more severe as the mixing time increases, since the anti-phase components are modulated by sin(πJtmix/2). This is clearly a problem for the measurements of weak dipolar couplings, which typically employ relatively long mixing times.

10742 J. AM. CHEM. SOC.

9

VOL. 124, NO. 36, 2002

The modified 3D TEDOR pulse sequences presented here make use of z-filter periods and band-selective refocusing pulses to suppress the ZQ and DQ spin coherences and refocus the J-couplings altogether, respectively. Both experiments generate pure absorption mode 2D spectra, where the cross-peaks located at (ΩS,ΩI) depend primarily on dipolar I-S correlations and can be used to determine internuclear distances. JA026385Y